Prediction of skin disease using bottleneck technology

The largest organ of the human body is the skin. As our skin protects internal organs and provides the main defense against UV light, it also gives space for microorganisms that cause skin diseases. It is critical for doctors to find different kinds of diseases just by a glance whereas computer does it with the help of deep research/learning. This paper depicts how to identify different kinds of skin diseases by just seeing the picture of the skin as an input, using transfer learning and a Bottleneck. The bottleneck is a layer preliminary to the final layer which does the classification. The output produced by this layer is adequate enough for the classifier to recognize the stage/classes to which the image belongs. The highlight of the bottleneck is, that there is no need for repeated deliberations required as every image is reprocessed multiple times during the training and bottleneck values are caught on disk. This method/system gives better accuracy than other algorithms, reduces the overall execution time and also displays the accurate result.